Think about all the things that need to happen for a human settlement to thrive: obtaining food, building shelter, raising children and more. There needs to be a way to divide resources, organize major efforts and distribute labor efficiently. Now imagine having to do this without any sort of planning or higher level communication. Welcome to the ant colony. Ants have some of the most complex social organization in the animal kingdom, living in structured colonies containing different types of members who perform specific roles. But although this may sound similar to some human societies, this organization doesn't arise from any higher level decisions, but is part of a biologically programmed cycle. In many species, all the winged males and winged virgin queens from all the nearby colonies in the population each leave from their different nests and meet at a central place to mate, using pheromones to guide each other to a breeding ground. After mating, the males die off, while females try to establish a new colony. The few that are successful settle down in a suitable spot, lose their wings, and begin laying eggs, selectively fertilizing some using stored sperm they've saved up from mating. Fertilized eggs grow into female workers who care for the queen and her eggs. They will then defend the colony and forage for food, while unfertilized eggs grow into males whose only job is to wait until they are ready to leave the nest and reproduce, beginning the cycle again. So how do worker ants decide what to do and when? Well, they don't really. Although they have no methods of intentional communication, individual ants do interact with one another through touch, sound and chemical signals. These stimuli accomplish many things from serving as an alarm to other ants if one is killed, to signaling when a queen is nearing the end of her reproductive life. But one of the most impressive collective capabilities of an ant colony is to thoroughly and efficiently explore large areas without any predetermined plan. Most species of ants have little or no sense of sight and can only smell things in their vicinity. Combined with their lack of high level coordination, this would seem to make them terrible explorers, but there is an amazingly simple way that ants maximize their searching efficiency; by changing their movement patterns based on individual interactions. When two ants meet, they sense each other by touching antennae. If there are many ants in a small area this will happen more often causing them to respond by moving in more convoluted, random paths in order to search more thoroughly. But in a larger area, with less ants, where such meetings happen less often, they can walk in straight lines to cover more ground. While exploring their environment in this way, an ant may come across any number of things, from threats or enemies, to alternate nesting sites. And some species have another capability known as recruitment. When one of these ants happens to find food, it will return with it, marking its path with a chemical scent. Other ants will then follow this pheromone trail, renewing it each time they manage to find food and return. Once the food in that spot is depleted, the ants stop marking their return. The scent dissipates and ants are no longer attracted to that path. These seemingly crude methods of search and retrieval are, in fact, so useful that they are applied in computer models to obtain optimal solutions from decentralized elements, working randomly and exchanging simple information. This has many theoretical and practical applications, from solving the famous traveling salesman problem, to scheduling computing tasks and optimizing Internet searches, to enabling groups of robots to search a minefield or a burning building collectively, without any central control. But you can observe these fascinatingly simple, yet effective, processes directly through some simple experiments, by allowing ants to enter empty spaces of various sizes and paying attention to their behavior. Ants may not be able to vote, hold meetings or even make any plans, but we humans may still be able to learn something from the way that such simple creatures are able to function so effectively in such complex ways.
想一下要發生多少事 才能讓一個人類聚落興旺: 覓食 建屋 生兒育女…等等 必須要有方法分配資源 安排主要工作 及有效分配勞工 想像一下在沒有任何計畫 或更高階的通訊下做這些事 歡迎來到蟻群世界 螞蟻在動物界中 具相當複雜的社會結構 生活在有架構的群體裡 裡面包含各式不同類型的成員 扮演特定的角色 雖然這聽起來近似於人類社會 但這種組織並不靠高階決策興起 靠的是生物程控周期 有許多品種的螞蟻 所有住在鄰近蟻群的有翅雄蟻 及有翅處女蟻后 每一隻都要從牠們的窩離開 到中央會面及交配 以費洛蒙引導彼此到繁殖地 交配之後,雄蟻一一殉難 而蟻后則試著建立一個新蟻群 能在合適地點定居下來的少數蟻后 會脫落翅膀 並開始下蛋 用牠們在交配時保存下來的精子 選擇性地讓蛋受精 受精卵長成雌性工蟻 負責照顧蟻后及牠所下的蛋 之後牠們要保衛蟻群 還要覓食 在此同時,未受精卵則長成雄蟻 牠們唯一的工作就是等待 直到牠們能離窩並繁殖 生命週而復始 那麼工蟻如何決定何時要做何事? 嗯,牠們其實並不做決定 雖然牠們沒有意向溝通的方法 螞蟻個體之間的確可以 透過碰觸、聲音及化學信號互動 這些刺激能完成許多事 例如在有螞蟻被殺時 能警示其他螞蟻 或是蟻后瀕臨生育年齡尾聲時發出信號 但是蟻群最令人敬佩的集體能力 是徹底和有效地探索大面積土地 不需任何事前計畫 大多種類的螞蟻視力很差 也只能聞到牠們附近的味道 再加上牠們沒有高階協調 這讓它們像糟透的探險家 但有一種了不起的簡單方法 讓螞蟻最佳化牠們的搜尋效率 就是以個體間的互動為基礎 改變牠們移動的模式 當兩隻螞蟻相遇 牠們藉觸角接觸感覺到對方 如果在一個小面積內有很多螞蟻 這會經常發生 使得牠們作出反應 以更迴旋、隨機的方式移動 使搜尋更徹底 但在大面積地區,螞蟻較少 這樣的碰觸較少發生 牠們能走一直線來涵蓋更多地面 當螞蟻以這種方法探索環境 牠們可能會碰到很多情況 像碰到威脅或敵人,或交替營巢處所 某些種類還有另一種能力,稱為招募 當這種螞蟻要找食物 牠會把食物帶回來 並在路徑上留下一種化學氣味 別的螞蟻就會循著這種費洛蒙路標 並在每次找尋食物的往返途中加重氣味 當那區的食物消耗殆盡 螞蟻就停止釋放回程的路標 這種氣味因而消失 螞蟻就不再被吸引至那條路徑 這種看來粗糙的搜尋及檢索方法 實際上非常有用 甚至可應用在電腦模式上 以獲得分散元件的最佳解 隨機工作與交換簡單資訊 這個方法有很多理論及實際的應用 從解決有名的銷售員旅行問題 到計算任務、最佳化網路搜尋 及賦能機器人群組搜尋布雷區 或燃燒中的建築物皆可 無需任何中央控制 但是你能直接觀察到這些 超簡單卻極有效的過程 只要透過幾個簡單的實驗即可 讓螞蟻進入不同大小的空間 並仔細觀察牠們的行為 螞蟻可能無法投票、開會 甚至做任何計畫 但我們人類仍可能從中學習 這麼簡單的生物